Introduction

Canine cervical soft tissue swelling in dogs is a common presentation through both primary care and emergency veterinary hospitals in the New England area. Differential diagnoses for cervical soft tissue swelling include abscesses, neoplasia, cysts, salivary mucoceles, and cellulitis.¹ To the authors’ knowledge, there are no publications of this scale exclusively describing cases of cervical abscesses (CAs) without identifiable foreign material. The current literature^2–6 describes CAs in dogs primarily in relation to identifiable foreign material, such as wood fragments, wound tracts, or gas pockets consistent with a penetrating injury or grass awns. Grass awns, although reported in cases of CAs, are more commonly associated with abscesses in alternate locations such as the external ear canal and interdigital webbing.^5,6 While the overall incidence of CAs secondary to penetrating foreign objects is not known, oropharyngeal stick injuries and foreign bodies of other etiology are considered top differentials.¹

Diagnostic imaging is indicated in cases of soft tissue swelling in the cervical region and is standard practice in human medicine.⁷ Various imaging modalities have been utilized to diagnose wooden and plant awn foreign bodies in dogs; CT and ultrasonographic imaging have been demonstrated to be the most useful.^4,8–12 Computed tomography imaging has been reported to have a sensitivity of 79% and specificity of 93% for wooden foreign bodies, with the ability to detect objects as small as 1 to 2 mm in diameter.⁴ Computed tomography has also been shown to be specific but poorly sensitive in identifying plant awn foreign bodies, as well as their migration tracts and other forms of vegetative foreign bodies.^12–15

Management of CAs involves a similar approach to that of wounds or abscesses within the subcutaneous or muscular layers in nearly any other location. Typical management involves surgical exploration of the abscess followed by flushing and debridement of the cavity. Closed-suction or Penrose drain placement is commonly employed.^2,13,14,16 In addition, broad-spectrum antimicrobial treatment is essential and ideally based on the results of culture and susceptibility testing.

In the authors’ experiences practicing in the Northeastern US, particularly in the Cape Cod region of Massachusetts, CAs in dogs (specifically in the ventral neck and intermandibular region) are a relatively common reason for presentation to veterinarians. Contrary to the publications^{1–6,9,11,13,14,16} regarding wooden foreign bodies or grass awns as inciting causes for these abscesses, the authors have rarely found foreign objects when no identifiable wound tract is present. Based on anecdotal experiential data, the authors suspected that these cases of idiopathic CAs follow a seasonal trend, occurring most frequently in the fall. These observations are in line with a recent publication¹⁶ based in the Greater Boston area. Additionally, anecdotal observations from regional colleagues in surrounding states support this phenomenon. Cape Cod is a popular vacation spot, attracting roughly 4 million visitors annually.¹⁷ The tourist season of Cape Cod spans from Memorial Day (the last Monday of May) through Labor Day (first Monday of September). As a result, the hospitals included in this retrospective study often have an increased overall caseload during the summer months, reflective of the influx of people with their pets to the area.

The purpose of this retrospective study was to describe a commonly encountered phenomenon of CAs in dogs without identifiable inciting cause in the Cape Cod region. Within this description, specific goals were to identify spatiotemporal patterns and potential environmental and dog-related risk factors (ie, regional land cover and sex, weight, age, and breed) associated with the development of these abscesses.

Methods

Electronic medical records from 2 specialty and emergency centers in the Cape Cod region (the only 2 in this geographical region) were searched for dogs presenting for possible cervical neck abscess from January 1, 2016, to December 31, 2020. Dogs included within the study (cases) were identified by utilizing search functions within the electronic medical record software. Searches were conducted with key terms in various fields in the database: abscess in the diagnosis field, surgical cervical neck explore in the treatment invoice field, and medical record entries including cervical abscess and neck abscess. Animals were excluded from the study if an abscess was not associated with the cervical region, if the record made note of identifiable external wounds that may have contributed to abscess formation, if the dog had a recorded history of chewing sticks, if the abscess was communicating with the oral cavity, or if an abscess was not identified or clinically suspected.

Each case’s medical record was reviewed; sex (intact male, intact female, neutered male, spayed female), weight (kg), age, breed, month and year of abscess diagnosis, US Postal Service ZIP code for the home address, clinical diagnosis, whether they received a CT scan or focal ultrasound imaging, whether they received aerobic/anaerobic bacterial cultures, culture results, surgical intervention, credentials of clinician performing surgical intervention, and antimicrobial utilized were all recorded. Age was evaluated as a continuous variable (in years, as integers). However, for those marked as < 1 year old, the following values were given: newborn to 2 months old, 0 years; 3 to 8 months old, 0.5 years; and 9 to 12 months old, 1 year.

The control group was created with an approximate 3:1 ratio of controls to cases, in order to improve statistical power, while balancing the necessary resources to capture the information. In order to create a control group representative of the true population, a list was generated including every dog seen during the timeframe of January 1, 2016, to December 31, 2020. There were 35,814 dogs seen between the 2 hospitals during the study period. The control group was created with a random number selector (Stat Trek; StatTrek.com) to include 204 control dogs.

To simplify the study and prevent potential masking of risk factors or confounders, no matching was performed for this case-control study. Only the first visit for a primary condition was allowed in the aleatory process, and the initial visit was used as the date for that individual in the study. There were no restrictions placed on the main presenting condition for control dogs; this resulted in controls being included from any of the specialty services (emergency, internal medicine, oncology, ophthalmology, and surgery). Once a control was selected for the study, sex, weight, age, breed, and US ZIP code were recorded.

Results

Sixty-seven dogs matched the inclusion criteria based on electronic medical record search from January 1, 2016, to December 31, 2020. Breeds were categorized based on the American Kennel Club’s 2023 list of recognized breeds. Any “boutique” breed that was not an American Kennel Club–recognized breed was categorized as “mixed breed.” Twenty-three mixed-breed dogs (33.8%) were reported. Other frequently recognized breeds included Golden Retriever (n = 8 [11.7%]), Standard Poodle (4 [5.8%]), German Shepherd Dog (3 [4.4%]), Labrador Retriever (3 [4.4%]), Beagle (2 [2.9%]), and Portuguese Water Dog (2 [2.9%]), and a collection of single cases of primarily large-breed dogs (22 [32.8%]). Of the breeds with CAs, only 1 was positively associated when compared to the controls. All 4 Standard Poodles in the study (combined cases and control) had a CA (P = .004; Fisher exact test). Interestingly, Labrador Retrievers were less likely to develop CAs compared to other breeds (OR, 0.23; P = .013; Fisher exact test; Figure 1).

View Full Size

Distribution of breed by cervical abscess status (top panel; cases = 67; controls = 204). Only breeds with multiple cases are presented, while the remaining breeds are grouped as “other.” The middle and bottom panels represent distributions of results from aerobic and anaerobic culture samples, respectively, collected from cases of idiopathic cervical abscesses.

Citation: Journal of the American Veterinary Medical Association 262, 10; 10.2460/javma.24.03.0228

• Download Figure
• Download figure as PowerPoint slide

The mean age at the time of presentation for CA was 5.5 years (range, 3 months to 15 years). Looking at unconditional associations (only the effect of age), younger dogs were found to be at a higher risk, with the odds of CA development reducing with age. For each additional year of age, the odds of having a CA reduced by 8.4% (OR, 0.916; 95% CI, 0.859 to 0.976; P = .007). Similarly, when assessing only the impact of weight on the odds of CA (unconditionally), weight was found to be a risk factor (linear increase) until 15 kg, where it plateaued afterward. More specifically, the odds of having CAs increased by 16.1% (OR, 1.161; 95% CI, 1.034 to 1.303; P = .011) for each additional kilogram until 15 kg (> 15 kg; OR, 0.983; 95% CI, 0.952 to 1.015; P = .296), at which point the risk remained elevated. It is worth noting that 36 controls did not have weight recorded; the majority of these were patients of the ophthalmology service. The median weight in CA cases was 24.4 kg (range, 5.9 to 59.5 kg). The median weight of control cases was 25.3 kg (range, 2.3 to 83.18 kg). Sex and neuter status were not associated with CAs (Table 1).

Cervical abscess location was recorded in all cases. The most frequently recorded location was “ventral cervical” in 32 cases (47.8%). Other recorded locations included “right ventral cervical” (n = 15 [22.4%]), “left ventral cervical” (13 [19.4%]), “intermandibular” (5 [7.5%]), “right submandibular” (2 [3.0%]), and “base of temporomandibular joint” (1 [1.5%]).

Bacterial cultures were submitted from 48 of the 67 cases. Aerobic culture was submitted from 45 cases. No growth was noted in 22 of these cultures (49%). The most frequently isolated aerobic organisms were β-hemolytic Streptococcus spp in 7 cultures (15.5%), Escherichia coli in 5 cultures (11%), Enterococcus spp in 5 cultures (11%), and Pasteurella multocida in 4 cultures (9%). Additional aerobic organisms include Klebsiella pneumoniae, Staphylococcus schleiferi, Streptococcus constellatus, Pasteurella gallinarum, Pantoea agglomerans, a Bacillus species, and an unidentified anaerobic organism. Anaerobic culture was submitted from 11 cases. Bacteroides spp were cultured in 6 cases (54.5%). Other anaerobic organisms cultured included Fusobacterium spp and an unidentified anaerobic species. No growth was identified in 3 cases (27%; Figure 1). Polymicrobial infections were identified, although uncommonly. Four cases had multiple aerobic isolates. Nine cases had both aerobic and anaerobic cultures submitted. A single case with both aerobic and anaerobic cultures submitted had no growth on either method; the remaining displayed bacterial growth. Of those cases, 3 (33%) had mixed aerobic and anaerobic bacterial species cultured.

Electronic medical records were reviewed to document typical treatment protocols. Of the 67 cases, 47 (70%) were treated surgically. Surgery was performed by board-certified surgeons in 34 of these 47 cases (72.3%). The remainder were performed by emergency associate veterinarians (6 of 47 [12.7%]), ACVECC residents (5 of 47 [10.6%]), and board-certified criticalists (2 of 47 [4.2%]). Surgical treatment uniformly consisted of abscess exploration, sterile lavage of the abscess, and a search to identify an etiology. No cases that involved surgical exploration documented identifiable foreign material, and most cases that involved surgical exploration specifically commented on a lack of identifiable foreign material. One case identified an amorphous necrotic tissue mass at the center of the abscess cavity. Most surgical cases involved some form of surgical drain placement. A Penrose drain was utilized in 36 cases (75.0%), whereas a closed-suction drain was placed in 1 case (2.0%). Surgery was performed without drain placement in 11 cases (22.9%). Abscesses were explored surgically and allowed to heal by second intention in 10 cases (20.8%). Wound exploration followed by management with a wet-to-dry bandage was utilized in 1 case (2.0%). Conservative management included abscess aspiration or antimicrobial treatment only. Aspiration of purulent fluid only (followed by oral antimicrobial treatment) was utilized in 16 cases (23.8%). Lone antimicrobial treatment was utilized in 3 cases (4.4%). Review of surgical versus nonsurgical treatment outcomes was outside of the scope of this study.

Regardless of whether a case received surgical intervention, all cases were treated with an antimicrobial. The most frequently utilized antimicrobial was amoxicillin–clavulanic acid in 59 cases (88.0%). One case was treated with meropenem in hospital but was de-escalated to amoxicillin–clavulanic acid following susceptibility testing. Enrofloxacin, clindamycin, cefpodoxime, doxycycline, and marbofloxacin were utilized in 3 cases (4.4%), 2 cases (2.9%), 2 cases (2.9%), 1 case (1.5%), and 1 case (1.5%), respectively. There were no cases that required alteration or prolongation of their antimicrobial protocols as documented in the medical record. No case was discharged with polypharmacy consisting of multiple concurrent antimicrobials.

Diagnostic imaging included CT and point-of-care ultrasound imaging. There were 6 cases (8.9%) that received CT prior to surgery. All 6 cases that received CT imaging underwent surgical exploration of the CAs. Two cases received CT imaging following surgery; one received imaging immediately postoperatively related to significant soft tissue swelling overlaying the trachea, which required temporary tracheostomy, and the second received a full-body CT 5 months after CA treatment for an unrelated issue. All CT scans were reviewed by a board-certified veterinary radiologist. No foreign material was noted in any case, including those with postoperative imaging. Point-of-care ultrasound of CAs was documented in 6 cases (8.9%), which was performed by a range of emergency clinicians, emergency and critical care residents, and board-certified veterinary criticalists. No visible or suspected foreign material was documented in any of these cases.

Calendar year was not associated with a significant difference in the number of cases (P = .516; Table 1). There was an overall significant seasonal effect (P = .006), with most cases occurring in September and October (P < .001 and P = .035, respectively). The ORs for September and October were 4.35 (95% CI, 1.97 to 9.61) and 2.56 (95% CI, 1.04 to 6.15), respectively. The space-time scanning statistic, carried out in SaTScan, detected a significant seasonal trend in the dataset, with the months of September and October having significantly higher risk than the rest of the year (P < .001). The distribution of CAs could not be explained by seasonal increase in case load alone. The busiest months for the hospitals (based on control data) were determined to be July and August, whereas the highest incidence of CAs was demonstrated to be in September and October (Figure 2). When comparing September and October to all remaining months, the OR was 4.03 (95% CI, 2.16 to 7.52), indicating that the odds of having an abscess in September and October were 4 times the odds of having one the remainder of the year (P < .001).

View Full Size

Histograms showing frequency of cases (top row) and controls (bottom row) across years, months, ages, and weights. Note that the x-axis labels “Year,” “Month,” “Age (years),” and “Weight (kg)” apply to the respective histograms in the top row as well. The x-axis labels for the Month versus Cases and the Month versus Controls histograms are abbreviations for the months of the year: January, February, March, April, May, June, July, August, September, October, November, and December.

Citation: Journal of the American Veterinary Medical Association 262, 10; 10.2460/javma.24.03.0228

• Download Figure
• Download figure as PowerPoint slide

There was a spatially dependent cluster of cases located between the 2 hospitals affecting 13 ZIP code areas (Figure 3). This cluster was identified by use of the space-time scanning statistic in SaTScan, where it was found that these 13 ZIP codes were at higher risk between August and November of 2019 (P = .043) than the rest of the region and time periods. No statistical significance was found for ZIP codes that were within a 70-km radius of either of the 2 clinics (P = .301; Table 1).

View Full Size

Cases of cervical abscesses (n = 67) in dogs and controls (204) in the Cape Cod area of Massachusetts, presented in 2 referral veterinary hospitals between 2016 and 2020. Subdivisions represent the US Postal Service ZIP code regions, and 2 cases had registered home ZIP codes outside of Massachusetts (one in North Carolina and another in Ohio). One case within Massachusetts is not shown since it was further north from Boston and outside of the map boundary. Seventeen controls had ZIP codes outside of Massachusetts (Texas [n = 1], Georgia [1], Virginia [1], Pennsylvania [1], New Jersey [1], Vermont [1], New Hampshire [2], New York [4], and Connecticut [5]), and 5 controls had ZIP codes within Massachusetts but outside of the map boundaries (not shown). A space-time scanning statistic, by use of a Bernoulli model, identified a high-risk spatial cluster comprising 13 ZIP codes (darker gray) between August and November 2019 (P = .043).

Citation: Journal of the American Veterinary Medical Association 262, 10; 10.2460/javma.24.03.0228

• Download Figure
• Download figure as PowerPoint slide

Further spatial investigations utilized Massachusetts land cover data to identify environmental risk factors. Unconditional logistic regressions identified the following land cover classes as potential risk factors (P < .15): pavement, railroads, forest, grassland, bare land, shrub swamp, marsh, pond, and coastal dune (Table 2). These analyses excluded dogs that had home ZIP codes outside of Massachusetts (effectively excluding out-of-state vacationers), which meant a total of 235 dogs were included in these analyses (cases, 61; controls, 174).

The final multivariable logistic regression (Table 3), accounting for multiple risk factors simultaneously, revealed the following significant risk factors: the month of September, the ZIP code cluster consisting of 13 ZIP code areas (Figure 3), age, weight, and the percent of land coverage within the same ZIP code area for “shrub swamp” and “pond” land cover. The odds of having a CA in the month of September were 4.8 (95% CI, 1.914 to 12.023) times greater than the odds of having one in the remaining months of the year, and the odds of dogs with a home address within the cluster of high-risk ZIP code areas having a CA were 4.4 (95% CI, 1.947 to 9.947) times the odds of dogs with home addresses outside of the ZIP code cluster having a CA. As dogs got older, their odds of CAs decreased by approximately 8.9% (OR, 0.911; 95% CI, 0.837 to 0.992) annually, and as dogs were larger (body weight), their odds also increased by 18.8% per kilogram (OR, 1.188; 95% CI, 1.053 to 1.341) until a maximum of 15 kg, where the odds plateaued. Dogs living in ZIP code areas with higher land cover from shrub swamp and ponds were at increased odds, with the ORs being 1.691 (95% CI, 1.225 to 2.333) and 2.017 (95% CI, 1.144 to 3.558), respectively, for every additional 1% coverage in a ZIP code area.

Discussion

This retrospective study evaluated CAs in dogs without an identifiable cause—without evidence of inciting foreign material and without clear communication with the oral cavity or the external environment. Idiopathic CAs are a common occurrence in the Cape Cod region of Massachusetts. Idiopathic CAs, as described by this retrospective study, are distinctly different from other common causes of cervical soft tissue swelling, such as neoplasia, salivary mucoceles, cellulitis, or abscesses attributed to wounds or embedded foreign material. This retrospective study evaluated the seasonality and risk factors associated with idiopathic CA formation, as well as provided an overview of diagnostic imaging and treatment options.

The tourist season of Cape Cod is considered to span from Memorial Day (the last Monday in May) through Labor Day (the first Monday in September) and is estimated to attract 4 million visitors annually.¹⁷ Electronic medical records were reviewed from two 24-hour emergency and referral hospitals—the 2 primary tertiary care facilities serving this region of Massachusetts. As such, these hospitals see an increased overall caseload during the summer months, with peak numbers in July and August (Figure 2).

It is unlikely that this population of dogs with CAs is related to other common causes of general abscesses, such as altercations with other dogs, wildlife, or wounds obtained from rough play outdoors. Efforts were specifically made to exclude cases that may have been attributed to these common causes by excluding patients with overt external wounds, draining tracts, or communication with the oral cavity. While it is possible that outliers in this group may have had an external penetrating wound that was not apparent at presentation, the authors feel that if this were the case for a significant number of dogs, the trends would be more in line with the seasonal population flux, which is distinctly not the case. The authors found that the incidence of CAs had a seasonal trend, with a peak in September and October. In addition, focal ultrasound and CT are reported to have reasonable sensitivity to detect not only plant material (grass awn) or wooden foreign bodies, but also their migration tracts, and none were noted in our observed cases through these imaging modalities.^4,8–12,14 However, these studies have not evaluated the ability of nonradiologist clinicians to detect these changes with ultrasound and no ultrasound was performed by a board-certified radiologist. Additionally, a minority of the patients included in this study had advanced imaging performed.

Our data demonstrated seasonal predilection, with 47.7% of cases being identified in the months of September and October. The spatiotemporal investigation and the separate land cover data analyses identified 1 region at the beginning of the Cape Cod peninsula that included 13 ZIP codes (Table 3; Figure 3) and 2 land cover classes (shrub swamp and pond) with higher risk of detecting dogs with CAs. Interestingly, dogs living in that high-risk area had odds that were approximately 4 to 5 times greater of developing CAs than dogs outside of this region, supporting the possibility of a common attribute in that region that may be responsible for these lesions.

While there is evidence to suspect geographical location and land cover class, the exact reason behind these findings remains to be determined. Despite there being no plant material identified in any of the cases in this study, the authors suspect that some form of plant or other organic material may be associated with CA development. This organic/plant material would likely be found in higher abundance in the region near the base of the Cape Cod peninsula and/or in plants growing in grasslands and near railroad tracks. In the current literature, when abscesses are related to wooden foreign bodies or plant awns, recovery of said foreign material is common. However, if the penetrating object were small enough to leave little entry tract and were more susceptible to destruction by the body’s natural defenses, it is possible that it could be decomposed by the time intervention is sought or evade visual detection. Introduction from this manner could involve external puncture or migration following ingestion. Although present in only a single case, the culture of P agglomerans provided some support for a possible organic/plant-related inciting cause. Pantoea agglomerans has been associated with penetrating trauma of plant material in both dogs and humans.^28–30 Further field investigations as to a potential plant-awn–like material present in the identified land cover types during the late summer and early fall season are required to better support this theory.

Breed was a risk factor in this study. Standard Poodles were the only breed found to be associated with CA formation. Interestingly, Labrador Retrievers appeared underrepresented. In general, medium- to large-breed dogs were found to be associated with CA formation. Given the similar body size of Standard Poodles and Labrador Retrievers and the apparent popularity of the Labrador breed amongst the control population, it is possible that differences in coat may be an associated risk factor.

Overall, younger dogs were predisposed to the development of CAs. This may be attributed to the curious and playful behavior of younger dogs. This finding is in line with studies^3,9 evaluating oropharyngeal injuries and grass awns in dogs, where the average age of dogs is typically 3 to 5 years.

Weight was found to be an important risk factor for CAs, with risk decreasing above 15 kg. This intuitively coincides with the reducing risk of CA formation with increasing age. The most common breeds identified to have CA formation were medium to large breeds. An inverse relationship between age and weight supports the idea that large-breed dogs that are curious and interactive with their outdoor environment are predisposed to CA formation. This theory also lends credence to the potential of a vegetative etiology of these abscesses that grows at a level aboveground that coincides with the neck region of these dogs; alternatively, there may be transoral migration of ingested vegetative material in these cases. The latter possibility, however, seems less likely given the specific association with large-breed dogs and the curiously lower risk in Labrador Retrievers, which are notoriously indiscriminate eaters.

Management of abscesses in veterinary medicine generally involves surgical exploration, lavage of the cavity, placement of surgical drains in cases with excessive dead space, and antimicrobial treatment. The management of this cohort of CAs mirrored these management strategies. Antimicrobials typically selected for dog abscesses include characteristics of broad-spectrum coverage and anaerobic activity, as was done for this cohort of cases. The majority of cases (88%) included in this study were treated empirically with amoxicillin–clavulanic acid while culture results were pending. In cases that had bacterial cultures submitted, none required alteration of this antimicrobial protocol following culture and susceptibility results. The most frequently cultured organisms in our cases included β-hemolytic Streptococcus spp, E coli, P multocida, and Bacteroides spp. This collection of organisms differs from bacteria commonly cultured from canine bites inflicted on humans, in which β-hemolytic Streptococcus spp, Bacteroides spp, P multocida, and E coli are relatively uncommonly isolated.^31,32 The common bacterial isolates in our study population did have significant overlap with the normal oral flora of dogs, although importantly, Moraxella spp and Pasteurella canis were not identified, which are commonly isolated from infected canine bite wounds.^32,33 The lack of positive culture for these species supports a non–animal-bite-wound etiology but also may be related to the small study population and limited number of cases where bacterial cultures were performed.

Among cases where bacterial culture was submitted, it was inconsistent whether anaerobic cultures, aerobic cultures, or both were submitted. “No growth” was reported more frequently in aerobic cultures versus anaerobic cultures (44% vs 27%, respectively). There was a single case where both aerobic and anerobic cultures were submitted with no growth reported. This individual was treated successfully with surgical debridement followed by amoxicillin-clavulanate. Potential explanations for negative culture include poor sample handling, unrepresentative culture sample, fastidious bacterial etiology, sterile disease process (sterile cellulitis), and undocumented preculture antimicrobial treatment. Based on this information, the authors recommend treatment submission of aerobic and anaerobic bacterial cultures on fluid or tissue on all patients and the use of amoxicillin-clavulanate as a first-choice empirical antimicrobial in similar situations.

While the literature surrounding CAs in dogs is relatively limited, a recent study¹⁶ published by an urban New England hospital had strikingly similar findings. The aforementioned study included 82 dogs with all-cause CAs that were treated surgically. Only 3.6% of the dogs were found to have an identifiable draining tract, and only 6% had identifiable foreign material. Additionally, that study found a similar seasonality with increased cases in August and September and very similar bacterial culture results (most frequent isolates including Streptococcus spp, Bacteroides spp, Pasteurella spp, and Actinomyces spp). The authors suspect that many of the cases from the aforementioned study may have a similar phenomenon to that documented in the present study, giving credence that this condition may be associated with a broader geographical range than the Cape Cod peninsula and, as such, is a valuable differential as a cause of cervical swelling.

There were several limitations to this study, related to its relatively small sample size, originating from 2 regional referral clinics, and the retrospective nature of the design. As this was an observational retrospective study, standardization of diagnostics and treatments was not possible. Not all patients underwent surgical exploration, and those that did were not all performed by a board-certified surgeon. Persistent foreign material may have remained within the abscess cavity in patients undergoing treatments that excluded surgical exploration. In isolated cases, an absence of grossly detectable vegetative foreign material has subsequently been identified on histologic assessment.¹³ It was not standard practice for the surgeons in this hospital to submit tissue in these cases for histopathologic evaluation. There was a variety of diagnostics involved. A significant limitation of this study was the lack of diagnostic imaging performed. Computed tomography and cervical ultrasound investigation were only performed in a minority of patients. Additionally, focal ultrasounds were not performed by a board-certified radiologist. Most cases in which point-of-care ultrasound was performed did not have documented cytology results, and bacterial culture was inconsistently performed. However, these cases were included on the basis of clinical diagnosis often based on compatible sonographic changes and frequent aspiration of grossly purulent fluid. These cases tended to be associated with significant owner financial constraints, and, as such, more definitive diagnostics had been declined. Many cases in the current study had cervical survey radiographs performed; these details were excluded in this study due to the known lack of sensitivity for detection of abscesses and organic matter foreign material. Had a higher proportion of patients undergone diagnostic imaging, it is possible that foreign material would have been identified more frequently among those cases, particularly in those patients where surgical exploration was not pursued. Despite the many cases with surgical treatment by board-certified veterinary surgeons and with a consistently available surgical record, some medical records may have been incomplete with missing documentation of external draining tracts or communication with the oral cavity. However, most surgical records performed by board-certified surgeons specifically commented on the lack of identifiable foreign material.

Although there was the occasional patient that returned to the clinic after the initial workup, for rechecks or suture removal, most patients had no follow-up information available in their medical record. Because of this, a follow-up analysis on complication rate or recurrence was not possible, meaning it was impossible to determine the efficacy of the described treatments and potential associations with recurrence. Persistent foreign material following surgery is a known cause of treatment failure.¹⁶ It is possible that abscess recurrence with nonidentified foreign material or antimicrobial failure occurred and that the authors have no documentation of it. For this reason, treatment efficacy was specifically not addressed within this manuscript.

To provide as true a representation as possible of potential risk factors, it was decided that all dogs presenting to the hospital from any service were to be included in the control group. Interestingly, 36 of the 204 controls (17.6%) randomly included in the study did not have recorded weights, and it was later found that patients visiting the ophthalmology service often did not have their current weight recorded during admission. Despite this potential selection bias, there were no other reasons to exclude any dogs from the control group. Importantly, by selecting controls only from the 1 regional referral hospital, the study findings may not necessarily extrapolate to all dogs in the general population of the Cape Cod region, since the controls represent a special group of dogs with existing conditions that require a visit to a veterinary specialist.

Although breed, sex, and weight were evaluated for significance, weights were not recorded for all patients in the control group. In total, weights were recorded for 168 of the 204 controls. It is possible that if weights were recorded for all these patients, a statistically significant difference may have been identified. Additionally, while weight was recorded for patients and some inference can be made as to general patient size, no metrics were available in relation to patient height. With a hypothesis related to potential plant or organic material, dog height may be related to an increased risk to thorns or awns of a particular plant in regard to the plant’s naturally found height (ie, “neck height”) or predisposition to getting caught in various types of furs (eg, long, curly, bristly).

Results of the present study showed that an idiopathic CA is an important differential for young, large-breed dogs presenting for evaluation of cervical swelling in the Cape Cod region of Massachusetts and may be a valuable differential in the greater Northeast region of the US. The development of idiopathic canine CAs in the Cape Cod region of Massachusetts has a strong seasonal trend with the highest odds of finding cases in late summer/early fall (September through October). Despite a high percentage of the patients in this study having surgical exploration of their abscesses, inciting foreign material was not identified in any of the cases presented. Risk factors that were identified to be associated with the development of CAs include young age, breed (Standard Poodle), weight, and ZIP code. Both the ZIP code areas located near the base of the Cape Cod peninsula and the amount and type of land cover found in the ZIP code area were associated with increased odds of developing CAs. Our findings raise the suspicion that plant material present in grasslands and near railroads may be in part responsible for the types of CAs described in this study; however, further investigation is required.

Supplementary Materials

Supplementary materials are posted online at the journal website: avmajournals.avma.org.

Disclosures

The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.

Funding

The authors have nothing to disclose.

ORCID

B. T. Welch https://orcid.org/0000-0002-6329-4216